Introduction

The nephrotic syndrome in children is characterized by the triad of heavy proteinuria (>1 g/m2/day or protein/creatinine ratio > 200 mg/mmol), hypoalbuminemia (<25 g/l) and the presence of generalized edema. So-called ”idiopathic” nephrotic syndrome (INS) is the most frequent cause and consists mainly of two histological subtypes, the minimal change nephrotic syndrome (MCNS) and focal segmental glomerulosclerosis (FSGS). Around 80–90 % of pediatric cases show minimal change disease on histological examination and most of these children respond to steroids (steroid-sensitive nephrotic syndrome, SSNS) as shown by the International Study of Kidney Diseases in Children (ISKDC), alleviating the need for an initial renal biopsy [1].

In contrast, the definition of steroid-resistant nephrotic syndrome (SRNS) by the ISKDC as persistent proteinuria after a 4-week course of oral prednisone (60 mg/m2/day) is still used in clinical practice. A renal biopsy is indicated in this situation and the majority of patients with SRNS will have FSGS, MCNS, or rarely diffuse mesangial proliferation (DMS). Treatment options in steroid-resistant nephrotic syndrome vary, but in the past few years considerable progress has been made by the use of calcineurin inhibitors and mycophenolate [2]. Unfortunately, a significant proportion of patients do not respond to conventional treatment, indicating the need for some alternative treatment approaches.

Treatment targeting B-cells has been used successfully in INS, e.g., oral or intravenous cyclophosphamide, but nowadays, specific B-cell-depleting antibodies are available, especially anti-CD20 (rituximab, RTX). Since the response to conventional treatment varies significantly between MCNS and FSGS, it is likely [3] that the pathogenesis of these disorders is also different (we will not discuss genetic causes of FSGS in this article). Therefore, MCNS and FSGS (including recurrent FSGS) and the discussion of the role of rituximab in the treatment of these disorders will be discussed separately.

MCNS: steroid-sensitive nephrotic syndrome

In 1974, Shalhoub [4] hypothesized that idiopathic nephrotic syndrome is a disorder of T-cell function because of the clinical association with Hodgkin’s disease, remission after measles infection, and for several other reasons. This was supported by many immunological findings [5] and also by the response to treatment with T-cell-specific immunosuppressants, e.g., calcineurin inhibitors. Recent data show that B-cell immunity is also altered in the idiopathic (mainly steroid-sensitive) nephrotic syndrome, e.g., persisting hypogammaglobulinemia in remission or an increase in the B-cell activation markers, especially in steroid dependency [6, 7]. Also, the therapeutic effect of immunosuppressants acting on B-cells (e.g., cyclophosphamide, partly mycophenolate) supports the role of altered B immunity in INS, although few studies have been performed yet in this field [8].

In our opinion, finding arguments for (or against) immunological treatment specifically for T- or B-cells in INS is difficult. First of all, despite the abundance of data, the exact immunological pathogenesis of INS is still unknown. Second, recent immunological studies show that a strict distinction of T- or B-cell responses is not useful, as there is a close network of immunological interaction between these two arms. Third, this applies to the action of immunosuppressive drugs, even if they target just one cell type. This is also true for anti-B-cell treatment with rituximab, because a variety of immunological (and non-immunological) mechanisms are involved, that go far beyond the circulating CD20+ B-cell (Fig. 1).

Fig. 1
figure 1

B-cells are multifunctional and regulate immune homeostasis in many ways; often, these effects are antibody-independent

Full size image

Rituximab effects: beyond the B-cell

B-cells do not only contribute to immune regulation by producing antibodies, but also by interacting with T-cells, producing regulatory cytokines and influencing apoptosis. It can therefore be assumed that the immune-modulatory effect of rituximab goes beyond simply reducing the number of antibody-producing and antigen-presenting B-cells. The extent and duration of rituximab’s action seem to be of great inter-individual variability, especially if concomitant immunosuppressive treatment is given. After a single dose of rituximab only B-cells in the peripheral blood, but not secondary lymphoid organs, are depleted and the functional properties of the remaining B-cells are changed [9]. Different mechanisms by which rituximab influences T-cell immunity in autoimmune disorders are discussed in the literature: reducing T-cell number and proliferation [10, 11] as well as modulating T-cell subsets [12, 13]. In patients with non-Hodgkin lymphoma, rituximab administration results in a transient, dose-dependent T-cell inactivation [14]. Therefore, the fact that rituximab is a B-cell-depleting agent, it exerts beneficial effects in patients with nephrotic syndrome, does not necessarily refute the T-cell hypothesis by Shalhoub. Studies explicitly examining the effect of rituximab on the T-cell compartment and its reactivity in children with nephrotic syndrome are needed.

It is well established that the B-cell activating factor of the TNF family (BAFF) rises in response to rituximab and this may contribute to its immune-modulatory effects [15]. Systematic investigations of serum BAFF in patients with nephrotic syndrome and correlation with disease activity and response to rituximab treatment may be of interest; in this situation, repopulation of B-cells may be very interesting. For instance, in children who received rituximab for acute rejection after renal transplantation, repopulating B-cells showed a mostly naive phenotype [16].

Use of rituximab in steroid-sensitive and -dependent nephrotic syndrome: from case reports to large series

The use of rituximab in steroid-sensitive nephrotic syndrome started after the initial report by Benz et al. [14], who used this drug in a patient with nephrotic syndrome who also developed idiopathic thrombocytopenia purpura (ITP) both went into remission with rituximab therapy after other treatments had failed. Several other case reports followed, but recently larger retrospective series have been published, which showed a beneficial effect in most steroid-sensitive patients [17–28]. Importantly, mainly patients with a refractory course were included in these studies, including patients who had relapsed and remained steroid-dependent, despite intensive maintenance immunosuppression. For an overview see Table 1.

Table 1 Rituximab in children with steroid- or cyclosporine-dependent nephrotic syndrome
Full size table

Retrospective series showed that patients were often able to reduce maintenance immunosuppression and some studies showed that patients achieved a full remission (varying between 25 and 83 %), often despite stopping all maintenance medication [17, 18, 25–28]. The international study group of Prytula et al. [21] demonstrated that maintenance immunosuppression including steroids could be reduced in most patients (86 %) following rituximab. Relapses after rituximab were related to the degree of B-cell depletion in some studies; however, some patients did not relapse, despite high CD19 counts [18, 26].

In addition to retrospective series the prospective study by Ravani et al. [23] documented the non-inferiority of rituximab compared with patients treated with prednisone and calcineurin inhibitors. Reduction of steroids was possible in almost all patients treated with rituximab, but not in the control arm, and relapses were significantly less frequent in the RTX group.

Kamei et al. [18] showed that a single dose of rituximab was able to initiate steroid-free remission in all patients. However, 75 % of patients relapsed and only 3 had sustained remission for more than 1 year. On the other hand, we reported results from a German registry and the duration of remission of patients receiving 1 or 2 vs those receiving 3 or 4 infusions of rituximab was not different [26].

While long-term remission after RTX with no further immunosuppression is the ideal outcome measure (cure of INS), steroid-sparing or reducing the number of relapses may also be deemed a success. Most patients treated with rituximab had a complicated (refractory), “difficult to treat” course and were exposed to a variety of long-term potentially toxic treatments. Previously, repeated cytotoxic treatments were used for these patients [29]. Thus, rituximab is at least an additional therapeutic option for patients with difficult-to-treat, steroid-sensitive nephrotic syndrome. Looking at the rituximab studies, however, one feels that outcome parameters should be more standardized and are often too descriptive.

Steroid-resistant nephrotic syndrome and recurrence of FSGS after renal transplantation

The pathogenesis of FSGS (except the genetic causes) is also not clear. Some data suggest a (soluble) host factor, e.g., produced by the immune system causing FSGS. This would explain, for example, the recurrence after transplantation; in this respect the sUPAR has recently become the most intensively discussed mediator [30]. The presence of a host factor is illustrated nicely by a case by Gallon et al. [31], where a living-related graft developed nephrotic syndrome after being transplanted into her brother with FSGS, but could be saved after being re-transplanted into another patient. Other studies look at structural changes, and just as an example one recent study showed increased expression of miR-193a in non-genetic FSGS, which inhibits WT1 expression [32]. Transgenic expression of the microRNA miR-193a in mice rapidly induced FSGS. As mentioned, treatment of primary non-genetic FSGS nowadays still includes immunosuppression; however, in addition to modulating the immune-response the effect of immunosuppressive treatment may be partially explained by a direct stabilizing effect on the podocyte, which has been shown nicely by Faul et al., especially for cyclosporine [33].

Rituximab and the podocyte

Non-immunological mechanisms of action on the podocyte have also been described for rituximab [34]. Fornoni et al. examined CD20 expression in kidney biopsies and suggested that rituximab has a therapeutic benefit through a non-immunological mechanism [34]. It is well known that rituximab can recognize CD20 on B-lymphocytes, but may also bind sphingomyelin phosphodiesterase acid-like 3b (SMPDL-3b) protein. Kidney biopsies from patients with recurrent FSGS after transplantation had less SMPDL-3b-positive cells per glomerulus compared with the stainings from nonrecurrent FSGS patients. In podocytes rituximab was able to prevent SMPDL-3b down-regulation. Overexpression of SMPDL-3b or treatment with rituximab was able to prevent the damage of the actin cytoskeleton in podocytes and reduced podocyte apoptosis caused by the patient’s sera. These remarkable data support a new model of an additional non-immunological mechanism for rituximab in (post-transplantation) FSGS.

Clinical results: rituximab in primary and recurrent FSGS

The response to rituximab in primary idiopathic FSGS or SRNS seems to be much more variable and less optimistic than in SSNS (Table 2). Initial reports were highly encouraging [35]; however, these results could not be confirmed by other series [20, 21, 26, 36–38]. An international registry [21] reported a superior initial response of steroid-sensitive (82 %) compared with steroid-resistant patients (44 %). In an American–Indian series by Gulati [20], extending their previous experience [35], response to rituximab was worse than initially described, but still 27.1 % reached full remission and 21.1 % partial remission.

Table 2 Rituximab in children with primary steroid-resistant nephrotic syndrome
Full size table

A recent open-label, randomized trial comparing two doses of rituximab with standard treatment with steroids and calcineurin inhibitors [38] did not show a reduction in proteinuria after 3 months. However, at a closer look 3 patients with “delayed-resistant” response entered remission with reduction of steroids and calcineurin; this also occurred in 3 “delayed-resistant” patients in the control arm. Inclusion of “delayed-resistant” or secondarily resistant patients may explain the difference between studies, since in the initial series by Bagga [35], 2 patients with initial steroid sensitivity were included and other patients had also been successfully treated with other drugs previously. As a further example in an unpublished series of 13 children with steroid-resistant nephrotic syndrome from Germany only 23 % showed a long-term benefit of rituximab, and again those with secondary steroid resistance responded better than those with primary resistance (Kemper, unpublished). Last but not least, one should be aware of potential reporting bias, because it is much more likely that unsuccessful cases will not be published.

Recurrence after transplantation

The use of rituximab in recurrent FSGS after transplantation has been summarized in a recent review [39] describing 39 patients, including 19 children. 64 % of patients achieved complete or partial remission often together with other treatment modalities, such as plasmapheresis. Young age and normal serum albumin levels were associated with a good response. Although a reporting bias for this subgroup of patients has to be anticipated, these data show that under certain circumstances rituximab is a therapeutic rescue option after recurrence of FSGS, which is a serious event. Further studies are necessary, e.g. relating to preemptive or very early treatment, e.g. when proteinuria is still very low indicating that podocyte damage is at an early, potentially reversible stage.

Open issues: rituximab in primary and recurrent FSGS

  1. 1.

    In general studies on the use of rituximab in FSGS (primary and after recurrence in the renal transplant) have to be viewed critically because of a publication bias, and in addition often other treatment modalities have been performed in parallel. It seems that individual factors related to initial presentation and previous treatment need to be characterized in a better way. Steroid resistance according to the ISKDC maybe not be exact enough since a patient with a late response or severe steroid dependency may be overlooked and therefore some authors recommend pulse steroids in the event of an initial nonresponse [40]. Also, a fraction of patients with FSGS respond to steroids (partially or completely) and may have a better prognosis. Thus, there is an overlap in steroid-sensitivity and resistance as well as FSGS and MCNS that needs to be considered when evaluating the effect of any treatment. Lastly, a fraction of patients have underlying structural (genetic) defects that respond poorly to the available treatment modalities.

  2. 2.

    Although treatment with rituximab in primary steroid-resistant nephrotic syndrome is not always successful, it may well be an option e.g. for late-responders or “delayed-resistant” patients as well as in patients who have gone into remission with standard treatment, but cannot be weaned or relapse on maintenance immunosuppression with rituximab, as indicated by the study by Bagga [35]. Also, optimal dosing, repetition of infusions, and other issues need to be addressed in prospective studies, although these may be difficult to perform because of the heterogeneity of these patients. It may well be that in FSGS timing (early vs rescue; preemptive in the patient after renal transplantation) of rituximab infusion (or other treatment) may well have an impact on treatment outcome.

Future use of rituximab: what about side-effects?

Briefly, although most patients seem to tolerate rituximab quite well [41], severe complications have been described, especially pulmonary complications [42], which can be fatal. In addition, progressive multifocal leukoencephalopathy (PML) has been reported after rituximab in patients with lupus nephritis. Too little is known about the long-term immunological complications after rituximab (hypogammaglobulinemia, effect on B- and T-cell function, response to vaccination, development of malignancy, etc.). Only if the risk/benefit profile proves to be equal to or even better than those of the drugs currently available, extending the indication for rituximab to less complicated patients is justified.

Summary

In summary, rituximab is an effective treatment option for children with steroid-sensitive nephrotic syndrome, but it is much less effective in primary steroid-resistant patients who are unresponsive to conventional treatment. In recurrent FSGS the effect of rituximab is not predictable; however, its use as a last resort to prevent graft loss may be an option in inducing complete or partial remission. Future studies should better define patient characteristics and entry criteria for use in SSNS and these studies need to address the effective dose, treatment modification, and ultimately the long-term risk profile. As yet, it is unclear how many patients reach long-term remission after this drug or become rituximab-dependent.